On 11/17/2016 11:51 AM, Andrew Sands wrote:

Hi all,

Just thinking I should do better with my data backup procedures. Currently throwing stuff onto portable USB drives in rotation - when I remember.

Is anyone using a 'cloud storage' provider and provide experiences either good, bad or indifferent.

And keeping on topic this is for my Linux boxen,

Cheers, keep steady and thanks in advance.

Andrew

Sorry for the late reply... was away doing stupid things in Queenstown over the weekend...

The problem with backing up to physical devices that you carry elsewhere is, of course that you forget, get them out of order and generally abuse them - eg cycling home in the rain - so that they get less reliable over time. The problem with backing up to the cloud is speed, especially if - like me - you live in the wops, and have a massive 100KB/s ADSL uplink speed or worse.

So, it's important to decide what you want the backup to do, as it'll colour what solution that you go for.

It's possible to mount an Amazon S3 bucket ( for example ) on a linux server, to provide a fairly cost-effective, remote target. There are a number of free products that make use of the '--link-dest' flag of rsync, as this will create a new directory tree for each backup, but will hardlink any files that are unchanged, dramatically reducing transfer and storage requirements ( well, after the initial upload ). I use a package called snapback2, but there are others out there. All the advantages of an incremental backup with none of the headaches (:

Obviously databases are harder to backup. Snapshots will never be able to take advantage of the above feature, but you can use log files, and weekly dumps to reduce volumes. It all just gets more complex, but if you need Point In Time Recovery, then you need to research the alternatives.

Encryption becomes less of an issue as your backups are an unmounted S3 bucket when not in use... worst case the keys are on the server, so your backups are only at risk if your server is breached.

The advantage of this approach is that it's easy to find and restore files: no need to download a full archive and extract stuff so in the case of an accidentally lost file, it's a trivial task to recover it... and you can do things to make it even easier like mounting the current backup as a readonly link for example.

( Yes, I know, all things are relative, and risk is a very subjective scale ).

hth,

Steve

--
Steve Holdoway BSc(Hons) MIITP
http://www.greengecko.co.nz
Linkedin: http://www.linkedin.com/in/steveholdoway
Skype: sholdowa

_______________________________________________
Linux-users mailing list
[email protected]
http://lists.canterbury.ac.nz/mailman/listinfo/linux-users

Reply via email to